A Methodology to Evaluate the Flail Space Model Utilizing Event Data Recorder Technology

Developed in the early 1980’s, the flail space model has become the standard method for estimating occupant risk in full-scale crash tests involving roadside safety features. The widespread availability of airbags and increased seat belt usage rates in today’s vehicle fleet, however, raise serious questions regarding the validity of the model. Recent implementation of Event Data Recorder (EDR) technology in a number of late model vehicles presents a different perspective on the assessment of the validity of occupant risk based on the flail space model. EDRs are capable of electronically recording data such as vehicle speed, brake status and throttle position just prior to and during an accident. Of particular interest is the EDRs ability to document the deceleration of a vehicle during a collision event. This paper presents a methodology utilizing EDR data to investigate the capability of the flail space model to predict injury to airbag-restrained occupants. Results of a preliminary analysis are presented based on implementation of the developed methodology on a limited data set. A majority of the analysis is limited to the occupant impact velocity due to complications in estimating the occupant ridedown acceleration. The longitudinal occupant impact velocity is found to be a good predictor of overall injury, chest injury and, to a lesser extent, lower extremity injury. For the head, and upper extremity body region, the longitudinal occupant impact velocity is a weak predictor of injury. Gabauer and Gabler 3 INTRODUCTION Full-scale crash testing has been the traditional method of evaluating the effectiveness of roadside safety improvements. The intent is to provide a measure of how a particular device compares to similar devices in representative worst-case impact scenarios. Currently, the evaluation criterion is based on structural adequacy of the appurtenance, post-impact vehicle trajectory, and occupant risk. As the purpose of roadside safety hardware is to be functional while minimizing the risk of occupant injury, the occupant risk criteria is vital to the assessment of these devices. For occupant risk determination in full-scale vehicle crashworthiness testing, crash test dummies have been developed specifically to mimic the human response in frontal and side impact collisions. Roadside hardware collisions, however, have a greater propensity for oblique impact angles. To date, no crash test dummies have been developed which can accurately reproduce the human response in this crash mode. Instead, the flail space model has been developed and implemented to evaluate occupant risk in roadside safety hardware crash tests. Evolution of the Occupant Risk Criteria Flail Space Model Prior to the introduction of the flail space model, a majority of the occupant risk criteria were based on limiting the peak 50 ms acceleration of the vehicle (1,2). Procedures for evaluating crash test data involved computing these values (both lateral and longitudinal directions) and comparing them with threshold limits. Although the determination of the occupant risk criteria was specified in excruciating detail, the authors of the guidelines cautioned that these evaluation procedures “...are not directly applicable to the complex highway collision” (2). In an attempt to better define the occupant risk criteria, Michie introduced the flail space concept formally in 1981 (3). This model was quickly incorporated into NCHRP Report 230, “Recommended Procedures for the Safety Evaluation of Highway Appurtenances” (4) and the subsequent update to these procedures, NCHRP Report 350 (5). Hypothesizing that occupant injury severity is a function of the velocity at which the occupant impacts the interior and the subsequent acceleration forces, Michie assumed the occupant to be an unrestrained point mass, which acts as a “free-missile” inside the occupant compartment in the event of a collision. The occupant is allowed to “flail” 0.6 meters in the longitudinal direction (parallel to the typical direction of vehicle travel) and 0.3 meters in the lateral direction prior to impacting the vehicle interior. Measured vehicle kinematics are used to compute the difference in velocity between the occupant and occupant compartment at the instant the occupant has reached either 0.3 meter laterally or 0.6 meter longitudinally. For ease of computations, the vehicle yaw and pitch motions are ignored, all motion is assumed to be in the horizontal plane, and the lateral and longitudinal motions are assumed to be independent. At the instant of occupant impact, the largest difference in velocity (lateral and longitudinal directions are handled independently) is termed the occupant impact velocity (VI). Once the impact with the interior occurs, the occupant is assumed to remain in contact with the interior and be subjected to any subsequent accelerations of the vehicle. The maximum 10 ms moving average of the accelerations subsequent to the occupant impact with the interior is termed the occupant ridedown acceleration. Again, the lateral and longitudinal directions are handled separately producing two maximum occupant ridedown accelerations. Both the VI and subsequent occupant ridedown acceleration are compared with established thresholds to ensure that the device does not create undo risk to the occupants of an impacting vehicle. Current threshold values as prescribed in NCHRP 350 (5), are summarized in TABLE 1. Note that this table excludes the limiting values for work zone devices and support structures. TABLE 1 Current Occupant Risk Threshold Values Occupant Impact Velocity Limits Component Direction Preferred Value Maximum Value Lateral and Longitudinal 9 m/s 12 m/s Occupant Ridedown Acceleration Limits Component Direction Preferred Value Maximum Value Lateral and Longitudinal 15 g 20 g Gabauer and Gabler 4 According to NCHRP 230 (4), the VI limit in the longitudinal direction was based principally on head impact experiments into windshields (6,7). The lateral threshold was based mainly on French accident statistics (8) and research aimed at developing FMVSS 214 (9), a U.S. vehicle standard for side impact protection. Occupant ridedown acceleration threshold values have been established mainly from exhaustive human impact tolerance review documents from the 1970’s (10,11). Although converted to SI units, the NCHRP 230 values were essentially retained in NCHRP 350 based on consultation with biomechanics experts, a General Motors (GM) research study (12), an evaluation report of current (NCHRP 230) guidelines (13), and an investigation of impact attenuator systems (14). Model Validity Unlike real-world crashes, roadside hardware crash tests are performed in a controlled environment with precise instrumentation. The result is exact values for the prescribed flail space model criteria. What is not well known is how this criterion relates to occupant injury in actual collision events. As discussed below, this relationship has traditionally been tenuous at best. Since decisions regarding the acceptance or rejection of roadside hardware are partially based on the flail space model criteria, there is a strong motivation to ensure the accuracy of this relationship. Stewart and Council (15) utilized accident data in an attempt to link occupant risk (as calculated in crash tests) to actual injury attained in collisions. The procedure matched instrumented full-scale crash tests with similar vehicle characteristics (make, model and year), crash characteristics (object struck, impact location on vehicle, etc.), and crash severity (as measured by vehicle deformation) in actual crashes. Results of this study indicated the lack of a strong relationship between injury severity and vehicle momentum change and 50-ms peak acceleration values. With regard to the flail space model, the limited data sample prevented any conclusions. In another study, Ray et al. (16) investigated the occupant injury mechanisms in longitudinal barrier collisions. The effort focused particularly on the lateral occupant impact velocity since a series of side impact sled tests, performed as part of the study, indicated that the current threshold might be overly conservative. By reconstructing seventeen longitudinal barrier accidents that produced severe occupant injury, the authors found that the lateral component of the first impact was not the cause of the serious injury in any case. A significant conclusion of this study is that the flail space model, although a useful tool for the estimation of occupant risk, does not appear to be a discerning factor in redirectional crash tests. Flail Space Model Revisited The original intent of the flail space model was to provide a gross indication of the likelihood of severe occupant injury to facilitate proper evaluation of roadside hardware devices. Despite the lack of evidence linking the flail space criteria to occupant injury, researchers have long questioned the validity of the assumptions of the model. As a result, several versions of the flail space model have emerged. In conjunction with the report evaluating NCHRP 230 procedures, Ray et al. (17) presented a method that considers the yaw rate of the vehicle using the coupled equations of motion. Similarly, Ross et al. (18) developed a program that considers vehicle yaw motion and a more exact “flail space”, depending on occupant seating location. Ray and Carney (19) developed a program that utilizes the coupled equations of motion, considers vehicle yaw motion, and calculates occupant position beyond the initial impact. Also, the European Committee for Normalization (CEN) has adopted a modified version of the flail space model to evaluate occupant risk (20). Differences include the use of resultant velocity and acceleration values to determine the impact velocity and ridedown acceleration, inclusion of vehicle yaw motion, and the addition of an acceleration-based model to account for belted occupants. Although the improved versions better characterize occupant motion, the linkage to occupant injury remains tenuous at best. Thus, it is questionable whether utilization of these improved models will impart a more efficient means of injury prediction. Advancing vehicle technologies and changing operating trends have further complicated the model assumptions and the correlation to occupant injury. At the inception of the flail space model in 1981, belt usage rates were approximately 15% and use of airbags was not widespread. Today, however, belt usage rates exceed 60% and driver and passenger airbags are required equipment on all new passenger cars. Is the unrestrained occupant assumption still valid despite these changes? Both previous investigations (15,16) of the flail space model were performed on a predominately non-airbag-equipped vehicle fleet and were unable address this issue. Also, these studies lacked vehicle kinematics information for the real-world collisions, which is crucial to the computation of the flail space criteria. The previously tenuous correlation to occupant injury coupled with fleet changes since the Gabauer and Gabler 5 inception of the flail space model raise serious questions regarding its current applicability and demonstrates the necessity for a reassessment of the model. EDR Technology and the Rowan University EDR Database Recent advancements in vehicle technology have allowed for an unprecedented opportunity to obtain information during a highway traffic collision. One such technology is Event Data Recorders (EDRs), which are being installed in numerous late model vehicles in conjunction with the advanced occupant safety systems. EDRs are similar to “black boxes” in airplanes as they record information in the event of a highway collision. Information typically stored by these manufacturer-specific devices includes seat belt status, deployment of the airbag, and vehicle speed prior to impact (21). Of particular interest to this study is the EDRs ability to record the vehicle velocity profile during a collision event. -35 -30 -25 -20 -15 -10 -5 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 Time (s) Ve lo ci ty C ha ng e (m ph ) FIGURE 1 Longitudinal Velocity Profile: 1999 Chevrolet Cavalier Under sponsorship of the National Highway Traffic Safety Administration (NHTSA), Rowan University is in the process of developing a first-of-a-kind database of EDR data collected from traffic collisions in the United States (22). Currently, the database consists of EDR data for over five hundred (500) cases, all of which are GM vehicles. These EDRs have the ability to store a description of both the crash and pre-crash phase of a collision. The crash parameters in the database include longitudinal velocity vs. time during the impact at 10 ms intervals (shown in FIGURE 1), airbag trigger times, and seat belt status for the driver. Pre-crash data includes vehicle speed prior to impact, engine throttle position as well as brake status for five seconds preceding the impact. As these cases were collected in conjunction with National Automotive Sampling System (NASS) studies, the corresponding NASS information is matched to the EDR data. NASS case investigators collect in-depth information about each crash including details regarding injury to the occupants. OBJECTIVE The purpose of this study is to present a framework for the use of EDR data to investigate the correlation between the flail space model and injury to airbag-restrained occupants. Results of a preliminary analysis are presented using the developed methodology on a limited data set. Gabauer and Gabler 6 METHODOLOGY The Rowan EDR database was first searched to identify those cases suitable for analysis. Suitable cases have the following characteristics: 1. Airbag deployment with associated velocity versus time data 2. Known NASS injury data for either the left or right front seat occupant 3. Comprised of a single impact only 4. Frontal collision In an attempt to utilize cases that have a higher potential for occupant injury, the data was narrowed to include only deployment events. Note that the typical velocity change threshold for airbag deployment in frontal collisions is approximately 5 m/s (11 mph) (23). Limited to information for a maximum of two impacts, the GM EDR will not capture all the events if a crash has more than two impacts. Reduction of the data set to include only single impact collisions ensures that the EDR velocity data corresponds to the injury-producing event. As the GM EDR only measures velocity information in the longitudinal direction, the data set has been constrained to frontal collisions only. A frontal collision, for the purpose of this study, is defined as damage to the front of the vehicle and a principal direction of force (PDOF) of 0 degrees plus or minus 10 degrees in either direction. A total of 88 cases have been identified as suitable for analysis; 68 left front seat occupant cases and 20 right front seat occupant cases. Seven (7) cases have been omitted from the analysis due to suspect velocity information. Note that there is potential overlap in the available cases. For instance, one vehicle may have injury indications for both left and right front seat occupants, resulting in two suitable cases for analysis. The final data set includes both frontal vehicle-to-fixed object (24%) and frontal vehicle-to-vehicle collisions (76%). If there is indeed a relationship between the flail space model and injury severity, it should be as equally relevant to vehicle-to-vehicle crashes as to vehicle-to-fixed object crashes. The following procedure was used to determine the longitudinal occupant risk criteria for the suitable cases in the EDR database: 1. Numerically integrate the longitudinal EDR relative velocity data to obtain occupant relative position as a function of time. 2. Interpolate to determine the time at which the occupant impacts the interior (relative distance = 0.6 meters). 3. Use the occupant impact time and the EDR relative velocity data to obtain the longitudinal VI. For cases where the theoretical occupant does not exceed the longitudinal flail space limit, VI is set to the maximum velocity change of the vehicle (as recorded by the EDR). 4. Obtain vehicle accelerations by numerically computing the derivative of the longitudinal EDR relative velocity and convert to G’s. 5. After the time of occupant impact, choose the largest absolute acceleration value as the occupant ridedown acceleration. 6. If the occupant does not reach the longitudinal flail space limit, the ridedown acceleration is set to zero. For cases where the occupant does not reach the flail space limit, NCHRP 350 specifies that VI should be set equal to the vehicle’s change in velocity that occurs during contact with the test article. As this would be extremely difficult to estimate with absolute certainty from EDR data, the maximum overall change in vehicle velocity is used to provide a conservative estimate of this quantity. A total of 32 of the 88 cases fall into this category; all have a VI less than 10 m/s and the highest injury level is AIS 3 (one case only). The occupant ridedown acceleration for these cases is set to zero since occupant impact does not occur. Note that NCHRP 350 provides no specific guidance on the occupant ridedown in this instance. Because the continuous longitudinal velocity profile is recorded at discrete points by the EDR, the question of validity arises for the computation of occupant ridedown acceleration (as it requires a derivative). To investigate the accuracy of the flail space model computations outlined above, six New Car Assessment Program (NCAP) frontal barrier tests were examined. Each car tested had GM EDR data available in conjunction with the more detailed vehicle acceleration data typically recorded for the test. Performing the flail space computations with each data source revealed an average error for the EDR-determined VI of 4 percent (6% maximum). The EDR consistently Gabauer and Gabler 7 overestimated the occupant ridedown acceleration on average by 40 percent with an overall range between 2 and 68 percent. Thus, this estimate of the occupant ridedown acceleration is an upper bound on this index, and should over predict injury potential. As long as the occupant impact occurs within the recorded velocity profile, VI will be accurate regardless of the completeness of the EDR velocity information. The ridedown acceleration, however, could be erroneous if the entire collision is not captured. The pulses of the 88 eligible cases have been examined for completeness (i.e. convergence to a constant velocity as shown in FIGURE 1). A total of 54 cases have been deemed complete; ridedown computations are limited to these cases only. For the quantification of occupant injury, the Abbreviated Injury Scale (AIS) is used as illustrated in TABLE 2 (24). The AIS scale is an injury severity metric that measures threat to life. The NASS/CDS data, collected in parallel with the EDR data, rates the severity of each occupant injury using this scale. Note that the intent of the flail space model threshold values is to prevent occupants from sustaining injury severity values of AIS 4 or greater (4). TABLE 2 The Abbreviated Injury Scale AIS Value Injury Characterization 0 No Injury 1 Minor 2 Moderate 3 Serious 4 Severe 5 Critical 6 Maximum/Fatal